Hire a specialist consultant (architect, security-auditor, business-analyst, performance-reviewer, refactor-critic, api-designer, data-engineer, qa-reviewer) to converse about the indexed grafel codebase. Interactive — one active consultant at a time, with Consult-Out to peers.
This repo is part of grafel group grafel. grafel is an architecture knowledge graph available via MCP. When you (an AI coding agent) need to understand how this codebase fits together, prefer the grafel MCP tools over grep + reading files.
Instructions for cajasmota/grafel, covering grafel — contributor agent guide, repo conventions, when this applies, when this does not apply and how to update.
Instructions for cajasmota/grafel, covering when to use grafel mcp vs grep, three concrete examples, grafel mcp, when to use grafel instead of grep and when grep is still better.
Add a new stack convention for the generate-docs skill. Run when the user has a repo whose framework or runtime is not covered by any existing file under skills/generate-docs/conventions/.
Teaches an AI agent to use grafel MCP tools as a structural second pair of eyes during code review. Covers impact radius analysis, orphan detection, dead-end flow discovery, cross-repo consistency checks, and pattern violation spotting — with a per-change-type tool sequence, a severity rubric, and worked examples.
Generate group-synthesised, PM-facing documentation — product capabilities, business domain glossary, user journeys as plain-language narratives, business rules reverse-engineered from code, and a business overview landing page. Zero internal symbols. Does NOT require tech docs to have been run first — graph-only…
Hire a specialist consultant (architect, security-auditor, business-analyst, performance-reviewer, refactor-critic, api-designer, data-engineer, qa-reviewer) to converse with you about the indexed codebase. Interactive, one consultant active at a time, Consult-Out to peers when needed.
Generate a privacy-preserving anonymized quality report for sharing with grafel maintainers, then layer an agent-driven "Findings & Interpretation" section on top that ranks the most likely extractor/resolver gaps, flags report self-inconsistencies, surfaces indexing phase timings, and gives maintainers a prioritized…
Annotate the graph with structured YAML frontmatter on httpendpoint, processflow, and messagetopic entities so the dashboard panels (Paths, Flows, Topology) light up. No prose docs — frontmatter only. Idempotent and delta-aware; safe to re-run after any index update.
Continuous health check of the grafel knowledge graph. Runs a head-to-head benchmark between grafel MCP and grep+read across 10-15 representative questions, judges quality against ground truth, and emits a shareable markdown report. Supports --since delta mode for CI gates and incremental re-runs.
Overview of the grafel skill family. Lists all skills with one-line purposes, shows the canonical execution chain, and suggests which skill to invoke first based on your goal. Does NOT run any analysis — purely informational.
Standalone agent skill for pattern discovery without doc generation. Use when the user wants a fresh pattern audit of an indexed group but does NOT want /generate-docs to regenerate prose.
Bidirectional sync between the per-group pattern store and a repo's version-controlled CLAUDE.md / AGENTS.md. Marker-wrapped, idempotent, and privacy-aware.
Surface and resolve residual edges — runtime dispatch, dynamic URLs, ambiguous bindings — so the graph is fit-for-purpose before any downstream skill consumes it. Lists pending repair candidates, auto-resolves unambiguous ones via templates, and walks the user through the rest interactively.
Two-phase security audit of the indexed group. Phase 1 is deterministic static analysis (runs inside the indexer via internal/security/). Phase 2 is LLM-driven semantic confirmation, ranking, and explanation — adds SecurityFinding entities to the graph and emits a security/ doc tier. Interactive by default; --auto for…
Generate the per-repo and group-level technical documentation set — overview, modules, reference, cross-cutting, group synthesis, cross-links, and patterns. For engineers. Runs 13 orchestrated passes using the grafel MCP as the primary navigation surface. Supports delta re-runs (only rewrite modules whose graph…
Test the grafel docgen LLM-mode loop end-to-end on a single entity. Used for iteration, smoke-testing, and verifying the emit→fill→apply roundtrip without manual command-construction.
Teaches an AI agent how to use grafel MCP tools effectively when working on a registered codebase. A task-oriented intent-to-tool map over the 22 canonical tools, plus orientation workflow and hard anti-patterns.